Yes, I suppose it's a performance wishlist bug. But really there is no reason for it to use stdio formatting; really the zero-padded case is no harder to do with direct ascii manipulation than the non-zero-padded case. (Also GCC should specialize and optimize the stdio calls, i.e., partially evaluate them, to eliminate the performance penalty of stdio routines. But let's not go there.)
> yes 10000000| tail -n 10000000 This wouldn't accomplish the desired task, wich is to produce output to write to a file. But even modified to do so, say by using head instead of tail, it wouldn't work for this particular use case. The idea is to write something unique to each block so the tricky capacity-faking controller cannot fool us by hashing together equal blocks, or whatever sneaky things they do. But this would work: $ yes | cat --number > /mnt/test-file & $ cmp /mnt/test-file <(yes | cat --number) So you're certainly correct, in that seq isn't really important here. I was just reporting a performance issue...
